Main idea

Writing code cost time, and that cost produced a side effect nobody had decided on: before going further, you often had to talk it through, arbitrate, have it reviewed, understand what you were doing. Understanding wasn't guaranteed by any discipline, it was imposed by the price of the gesture.

When a few hours are enough where it used to take several days, that de facto guardrail disappears — and it leaves nothing behind, because it had never been instituted. You can then move faster than your own understanding, on screens that respond, a database that exists and an API that runs, with no moment forcing you to know why things are done that way.

So the missing discipline isn't a lost discipline: it is a discipline that never had to exist, and that has to be written for the first time.

Why it matters

This explains why even the most experienced teams are caught out. They haven't loosened their practices; they have lost an involuntary mechanism their practices leaned on without knowing it.

It also points to the remedy: not going back to the old practices, but making explicit — mandatory review, written architectural decisions, an understanding budget per feature — what slowness produced for free.

Nuances and limits

Outrunning your own understanding isn't always a flaw: it is exactly what you are after when exploring, when you want to see what a solution would look like before knowing whether it is worth anything. The danger begins when exploration turns into production without anyone acknowledging the crossing.

And slowness imposed the conversation without ever guaranteeing its quality: plenty of slow systems were badly understood all the same.

Open questions

• By what signal does a team recognize that an exploratory prototype has just been promoted to production without an explicit decision?